“ Why Generalists Triumph in a Specialized World ”

by David Epstein

Book Cover

  • Rating: 3.75 / 5

I liked this book, even if it’s not super practical, it’s thought provoking, which I guess is the intent.

Summary

Two kinds of environment: kind and wicked. Kind has clear rules and repeatable patterns (chess, tennis), wicked has no clear rules, complex arrangements, and is hard or impossible to predict. Most of the world is wicked.

The orthodoxy wants that people get specialized and start specialization young. However, evidence shows that generalization provides more ideas and better results in wicked environments.

Some strategies and benefits of generalist thinking

  • Analog thinking: use experiences from other domains to realize connections
  • Diversity of experience: curiosity and knowledge on many domains enables analog thinking
  • lateral thinking with withering technology: reuse mastered technology in novel ways instead of racing to innovation through brute force
  • see beyond the limited view of expertise when analyzing a problem, and notably, leveraging views from more people, including from non-experts, or less experienced people, which increases the range of ideas at disposal
  • abandon familiar tools in unfamiliar situation, be flexible rather than following the process blindly. Thinking of other domains.
  • Deliberately experiment with other domains, or out of the standard process, to stimulate ideas and discoveries.

Accept some uncertainty, detours, exploration, as an investment to discover. If you want the sky highs, you have to tolerate a lot of lows.

Notes

Two types of education - specialized (e.g. Tiger Woods who trained from a very early age on just golf) and generalized (e.g. Federer who trained in a large variety of sports including a ball and developed a range of corresponding skills)

Two types of learning environments - kind where things are somewhat predictable, and deliberate practice brings recognition of patterns and better results. Wicked, where rules are not clear, complex behavior is less predictable, and experience brings limited certain outcomes.

In kind environment (e.g. chess, tennis, golf), excellent tactics (i.e. raw skills) + basic strategy bring you a long way. In wicked environments, they matter less. Most environments are wicked

  • Automation is excellent at tactics, but less good at strategy. Humans are good at strategy from variety of experience. Combination of automation to handle tactics and humans to determine strategy brings excellent results
  • learning in a kind environment (specialized) can in fact impede adaptability if one of the variables changes.

Specialized training in kind environment produces narrow skills. Training in more areas allows recognizing different patterns and more adaptability. these are intuitive experts. They recognize patterns, then can think about them, but usually intuition is good

  • specialization looks at the detail, but you also need generalization to look at the system

Fermi approximations: approximations based on little data you know (e.g. How many Piano tuners in Chicago)

Learning:

  • Most important is struggle. Struggle is a sign of learning
  • Spaced repetition works best for long term learning, and also interval (vary exercises rather than repeat the same over and over again.)
  • Understanding the underlying concepts rather than just procedure.

Analog reasoning is applying analogies from other fields to a problem. To be able to reason analogly requires a diversity of experience and learnings.

  • we tend to confuse details with certainty, and estimate over optimistically what inside a domain compared to what they are outsiders of.
  • when looking for answers in a field we’re an expert of, tendency is to search within the field rather than diversify to other fields.
  • look for similarities with other domains

Diversity of experience is necessary to understand what we like. Diversity of training and jobs is desirable before committing. Advice against quitting something that does not fit with us is common and wrong. Test and learn, rather than plan and implement. Instead of working towards a goal, work from a promising situation. Avoid sunk cost fallacy.

Careers tend to go to specialization, however the concentration of experts means that all questions answered within the fields are answered, and the questions from out of the field are not.

Lateral thinking with withering technologies is looking for new uses of well understood technologies. It allows escaping the brute-force race to always try and find newer technologies, which has a lot of competition, and instead focus on actually fixing a problem (think: Nintendo).

  • don’t discard seemingly trivial ideas
  • think of related experiences in other domains

Beware of expertise

  • In prediction, experts often get beaten by crowd-sourcing using people with broad interests. When situation is sufficiently complex, expertise brings little value in predicting.
  • Experts are prone to confirmation bias.
  • When predicting complex and wicked situations, look at situations with a general similar structure rather than relying on details and personal expert opinions. (example: will Greece leave the EU - look at other failed international negotiations)

In the face of unfamiliar challenge, don’t hesitate to drop your familiar tools. Extend the set of data you consider, look for approaches you didn’t consider, change the way you process data.

  • beware of interrupting the chain of communication because of the chain of command
  • beware of usual, narrow processes in novel situations
  • speak up to contribute to diversity of ideas and opinions, even if data is missing, or against the protocol
  • hear everybody’s idea

Deliberately reserve time for experimentation and doing different things or things in different ways

  • standardization of how to do things slows innovation
  • reading outside of your domain helps making connections
  • curiosity on random, trivial things, led to later, unrelated discoveries. E.g looking at retroviruses in dogs before the first known retrovirus in human (hiv) saved decades of development to find treatment
  • take the time, chat, bounce ideas

Accept some uncertainty, detours, exploration, as an investment to discover. If you want the sky highs, you have to tolerate a lot of lows.


About Reading Notes

These are my takes on this book. See other reading notes. Most of the time I stop taking notes on books I don't enjoy, and these end up not being in the list. This is why average ratings tend to be high.